The transient supersaturation in a system undergoing Ostwald ripening is related to the cluster formation energy $E_{\mathrm{fc}}$ as a function of cluster size $n$. We use this relation to study the energetics of self-interstitial clusters in Si. Measurements of transient enhanced diffusion of B in Si-implanted Si are used to determine $S\left(t\right)$, and inverse modeling is used to derive $E_{\mathrm{fc}}\left(n\right)$. For clusters with $n>\mathrm{}15\mathrm{}$, $E_{\mathrm{fc}}\approx 0.8\mathrm{eV}$, close to the fault energy of $\{113\}$ defects. For clusters with $n<\mathrm{}10\mathrm{}$, $E_{\mathrm{fc}}$ is typically 0.5 eV higher, but stabler clusters exist at $n\approx 4$ ( $E_{\mathrm{fc}}\approx 1.0\mathrm{eV}$) and $n\approx 8$ ( $E_{\mathrm{fc}}\approx 0.6\mathrm{eV}$).

We demonstrate that the depth distribution of defects in MeV implanted n-type and p-type crystalline Si is severely affected by the impurity content of the material. Silicon samples with different concentrations of dopants (P or B) and intrinsic contaminants (i.e., C and O) were implanted with 1 or 2 MeV He ions to fluences in the range 2.5×10⁸–1×10¹³/cm². Using deep-level transient spectroscopy and spreading resistance measurements, we have identified the defects and determined their concentration and depth distribution. It is found that less than 4% of the defects generated by the beam escape recombination and are stored in electrically active, room temperature stable defect clusters, such as divacancies and carbon–oxygen pairs. When the concentration of these defects is much smaller than the doping level, their profile mirrors the initial defect distribution, as calculated by transport of ions in matter (TRIM), a Monte Carlo code. In particular, the profile presents a maximum at the same depth predicted by TRIM and a width which is strongly dependent on the impurity content of the substrate. Indeed this width can be as large as 2 μm when implants are performed on a lightly doped, pure epitaxial substrate and returns to the value predicted by TRIM (∼0.5 μm) upon increasing the concentration of dopants and intrinsic contaminants which act as traps for the diffusing point defects. The broadening of the concentration profiles is however shown to be unavoidable at high implantation fluences when most of the traps are already full and unable to interrupt the free migration of newly generated defects. Finally, by comparing defect distributions in n-type and p-type samples we have detected the spatial separation between vacancy-type and interstitial-type defects, resulting from the ion momentum transfer. The observed phenomena are explained in terms of a trap limited diffusion of the defects generated by the beam. These effects are only observable for a light ion such as He since direct defect clustering within the diluted collision cascades is expected to be significantly inhibited.

Excimer laser annealing (ELA) of ultra-low-energy (ULE) B-ion implanted Si has been performed. High-resolution transmission electron microscopy has been used to assess the as-implanted damage and the crystal recovery following ELA. The electrical activation and redistribution of B in Si during ELA has been investigated as a function of the laser energy density (melted depth), the implant dose, and the number of laser pulses (melt time). The activated and retained dose has been evaluated with spreading resistance profiling and secondary ion mass spectrometry. A significant amount of the implanted dopant was lost from the sample during ELA. However, the dopant that was retained in crystal material was fully activated following rapid resolidification. At an atomic concentration below the thermodynamic limit, the activation efficiency (dose activated/dose implanted into Si material) was a constant for a fixed melt depth, irrespective of the dose implanted and hence the total activated dose was raised as the implant dose was increased. The electrical activation was increased for high laser energy density annealing when the dopant was redistributed over a deeper range.